Machine learning
Jot — Offline Notes with a RAG agentJEm — Offline Notes with a RAG agent
A note-taking app where an Ollama-powered agent answers questions about what you've written, entirely offline via retrieval-augmented generation.
Role
Solo · design + implementation
Timeline
2024
Stack
TypeScriptTauriPythonOllamaRAG
Overview
Jot pairs a normal note-taking interface with a local agent: instead of exact-match search, you can ask it questions and get answers grounded in your own notes.
A Python RAG backend handles retrieval over your notes and feeds relevant context to a local Ollama model, so the whole thing runs offline with no cloud calls or API keys.
What I built
01
Fully local pipeline — Ollama serves the LLM and a Python service handles retrieval, so notes and queries never leave the machine.
02
Electron/Tauri frontend gives it a native note-taking feel while the RAG agent runs as a background service.
03
Ask-your-notes interface: query in natural language and get an answer synthesized from the relevant notes, not just a list of matches.
Next project
UCR Map — campus navigation app
→